1.1 --- /dev/null	Thu Jan 01 00:00:00 1970 +0000
     1.2 +++ b/src/SFMT/CHANGE-LOG.txt	Sat Mar 03 10:31:27 2012 -0600
     1.3 @@ -0,0 +1,55 @@
     1.4 +ver 1.3.3
     1.5 +-------
     1.6 +change condition compile of do_recursion in SFMT.c
     1.7 +
     1.8 +ver 1.3.2
     1.9 +-------
    1.10 +bug fix to_res53_mix and genrand_res53_mix.
    1.11 +
    1.12 +ver 1.3.1
    1.13 +-------
    1.14 +gcc compile option changed form -O9 to -O3.
    1.15 +add functions genrand_res53_mix and to_res53_mix.
    1.16 +bug fix about definition of ALWAYS_INLINE.
    1.17 +add new definition PRE_ALWAYS for MSC.
    1.18 +
    1.19 +ver 1.3
    1.20 +-------
    1.21 +bug fixed: -DONLY64 without -DBIG_ENIAN64 had been generating
    1.22 +wrong sequence.
    1.23 +bug fixed: There is no documentation about BIG_ENDIAN64.
    1.24 +add automatic endian check by __BIG_ENDIAN__ predefined macro.
    1.25 +bug fixed: change == in check.sh to =
    1.26 +add SFMT-params216091.h
    1.27 +add AltiVec parameter format for systems which are not osx.
    1.28 +change Makefile for systems which are not osx and support AltiVec.
    1.29 +change sample2 of howto-compile for Free BSD.
    1.30 +change source files for BORLANDC and Visual Studio.
    1.31 +change period certification code more smart.
    1.32 +add params directory.
    1.33 +
    1.34 +ver 1.2.1
    1.35 +-------
    1.36 +Fix typo in SFMT-alti.c SFMT-sse2.c
    1.37 +marge SFMT-alti.c and SFMT-alti.h into SFMT-alti.h
    1.38 +marge SFMT-sse2.c and SFMT-sse2.h into SFMT-sse2.h
    1.39 +This version is not released.
    1.40 +
    1.41 +ver 1.2
    1.42 +-------
    1.43 +Support many periods: 2^{607}, 2^{1279}, 2^{2281}, 2^{4253}, 2^{11213},
    1.44 +2^{19937}, 2^{44497}, 2^{86243}, 2^{132049}
    1.45 +Fix typo in LICENSE.txt.
    1.46 +Add cast to vec_perm for SFMT-alti.c, SFMT-alti64.c.
    1.47 +combine source codes.
    1.48 +
    1.49 +ver 1.1
    1.50 +-------
    1.51 +The period certification method is changed from constant to function.
    1.52 +The convert functions from 32-bit and 64-bit integer to double are added.
    1.53 +The documentation is changed.
    1.54 +Sample programs are added.
    1.55 +
    1.56 +ver 1.0
    1.57 +-------
    1.58 +The first version.